DataHub OSS Core is open and capable, but production self-hosting is a real platform-engineering project: Kafka, Elasticsearch, a metadata store, the GMS backend and a separate frontend, all stood up, tuned and maintained by your team. Dawiso is managed in every deployment option - SaaS, your Azure, or on-prem - with the full catalog on every plan and no self-hosted infrastructure to operate.
TL;DR
No DevOps team required. The full catalog on every plan, no Cloud-gated lifecycle. Managed agents on day one, MCP from Corporate.
Self-guided, in the real product. No sales call; a work email unlocks the full tour.
Moved from DataHub to Dawiso
The close dialogue with the Dawiso team is what made the pilot work: when our model didn't fit the tool, they adapted the tool to us.
Daniel FalleniusInformation Architect, Bertel O. Steen
~60 daysfrom kick-off to a fully operational solution, DataHub content migrated at Bertel O. Steen
Built on customer feedback and publicly available sources.
Features
Deployment and operations
Managed in every option (SaaS, your Azure, or on-prem) with 100% feature parity across deployment options - no self-hosted infrastructure to run
OSS Core is a self-hosted multi-service stack (Kafka, Elasticsearch, metadata store, GMS, frontend) needing a dedicated platform team; SLA-backed 99.5% availability and a dedicated customer success team are positioned as DataHub Cloud
Time to value
Fully managed; first use cases in 2-4 weeks; full named rollout ~11 weeks (P3 Parks, 78 days)
4-12 weeks with a strong engineering team; longer as adoption and governance are on you
Open-source story
One product, one feature set, no upgrade tax
OSS Core is genuinely open; certification workflows, governance automations, shift-left governance positioned as DataHub Cloud
Governance workflows
Approval flows, lifecycle states, Excel import/export included
Governance-lifecycle tooling (certification workflows, no-code automations) is positioned as DataHub Cloud
AI agents and MCP
Managed production MCP read+write (Databricks MCP Marketplace) from the Corporate plan, Technical Writer Agent and Agentic Data Governance Steward in every plan
DataHub's MCP server (acryldata/mcp-server-datahub) is self-hosted - you run and maintain it against your own instance; Cloud adds managed MCP and advanced agentic features
Total cost trajectory
Predictable per-seat, the full catalog in every plan
OSS free but engineering-heavy; Cloud pricing is opaque (Enterprise on request)
02Customer proof
From open-source DataHub to a production-ready catalog in about 60 days
Bertel O. Steen ran its catalog on open-source DataHub. Dawiso migrated the business glossary and data product documentation and had a customized, production-ready environment in place in roughly two months.
~60 daysfrom kick-off to a fully operational solution, DataHub content migratedBertel O. Steen
-OSS version missing Excel import/export for term curation
-Approval workflows and lifecycle states sit in DataHub Cloud
-Cross-domain term modelling requires custom YAML and ingestion code
GlossaryOSS GAP
Customer LTV
⚠Excel I/O missing in OSS
Approval flowCloud only
LifecycleCloud only
Cloud upgrade required
AIProduction agents on day one, MCP from the Corporate plan
AI
AI Context Layer
Production agents on day one, MCP from the Corporate plan
dawisoWinner
Native MCP + agentic stewards
✓REST API in every plan, MCP server from the Corporate plan
✓Agentic data steward + technical writer included
✓No separate vector store, no Cloud upgrade required
Where's revenue?
AI · 96%
fct_orders_daily · Finance · 04:12
Ask anything…›
DataHubLimited
OSS MCP shipping; managed agentic features Cloud-gated
-DataHub OSS ships acryldata/mcp-server-datahub on GitHub (self-hosted, fully featured)
-DataHub Cloud adds managed MCP and advanced agentic features
-No pre-built Technical Writer Agent or Agentic Data Governance Steward in OSS or Cloud
OSS MCP + Cloud agentsCLOUD-GATED
Where is revenue defined?
🔒
OSS MCP self-hosted; managed agentic in Cloudacryldata/mcp-server-datahub on GitHub
OSS MCP self-hostManaged MCP & agents Cloud-only
LINEAGELive and editable, not batch-ingested
LINEAGE
Live data lineage, editable in-product
Live and editable, not batch-ingested
dawisoWinner
Auto + manual edits, column-level, real-time
✓Live lineage from warehouse and transformation tools
✓Edit edges in-product where the parser misses
✓Impact analysis on every schema change
LINEAGE · COLUMN-LEVEL
4 nodes · 4 edges · realtime
DataHubLimited
Ingestion-driven lineage with manual gaps
-Lineage depends on recipe ingestion runs, not real-time observation
-Manual lineage edits require custom emitter code, not in-product fixes
-Column-level coverage uneven across BI and transformation tools
LineageINGESTION RUN
⚠YAML emitter edits only
Last run6h ago
Manual editsvia code
BI coverageuneven
Recipe-drivenNo in-product edit
DEPLOYMENTManaged and complete, not a stack you operate
DEPLOYMENT
Managed catalog vs a self-hosted platform you run
Managed and complete, not a stack you operate
dawisoWinner
Delivered managed in every deployment option
✓Managed in every option: SaaS, your own Azure, or on-prem - no self-hosted infra to operate
✓One product, one data model: business and technical metadata linked by design, with the same features on SaaS, Azure and on-prem
✓Account manager from the Corporate plan - customer success isn't paywalled to an enterprise tier
✓In-product customization of schemas, workflows and roles - no services engagement, and fast time-to-value with 40+ connectors (2-4 weeks)
LINEAGE · COLUMN-LEVEL
4 nodes · 4 edges · realtime
DataHubLimited
OSS Core is open, but production self-hosting is a platform-engineering commitment
-Production self-hosting stands up Kafka, Elasticsearch, a metadata store, the GMS backend and a separate frontend - reviewers note it needs significant engineering investment for setup, tuning and maintenance
-Realistically viable with a dedicated platform team running Kafka/Elasticsearch/Kubernetes (production guidance: 16GB+ RAM, 4+ cores on dedicated infra)
-Governance-lifecycle tooling - certification workflows, governance automations, shift-left governance, no-code automations - is positioned under DataHub Cloud
-SLA-backed 99.5% availability and a dedicated customer success team are positioned as DataHub Cloud; OSS Core users rely on community support and carry their own uptime
LineageINGESTION RUN
⚠YAML emitter edits only
Last run6h ago
Manual editsvia code
BI coverageuneven
Recipe-drivenNo in-product edit
PACKAGINGOpen about packaging, not open-source theatre
PACKAGING
Procurement model
Open about packaging, not open-source theatre
dawisoWinner
One product, the full catalog in every plan
✓No OSS-to-Cloud upgrade tax for governance features
✓AI and MCP included, not a Cloud-only add-on
✓Predictable per-seat, no per-table metering
PRICE TIER
dawiso$
DataHub$$
DataHubLimited
OSS framed as open, Cloud holds the governance
-OSS version missing Excel import/export, lifecycle, approval workflows
-Self-hosting OSS is engineering-heavy (Kafka, Elasticsearch, GMS, frontend)
-DataHub Cloud bills per-table at scale, with advanced AI behind the same gate
PlansOSS / CLOUD
OSS self-host engineer-heavy
Cloud entry 🔒 sales
Cloud enterprise 🔒 per-table
See the difference yourself. Click through the real product, at your own pace.
04Price and ratings
What it costs, and what users say
A 0-10 total-cost score from public AWS Marketplace starting prices, next to verified G2 ratings.
Total cost of ownership score, 0-10. Lower is better.
Dawiso
2.1/10
Collibra
9.8/10
Ab Initio
6.8/10
Atlan
5.6/10
Select Star
5.0/10
DataHub
4.1/10
OpenMetadata
4.1/10
Alation
3.2/10
DataGalaxy
2.7/10
Secoda
2.6/10
TCO scores anchor on public AWS Marketplace starting subscription pricing, normalized 0-10 where lower is better. Full 3-year program cost varies by services SOWs, paid modules (DQ, lineage, AI agents), and steward FTE. Methodology updated 2026-05.
CRITERIA
Ease of Use
9.7
8.6
Ease of Setup
9.4
7.8
Quality of Support
9.4
8.4
Data Lineage
9.3
8.7
05FAQ
Frequently asked questions
Still deciding? Contact us and we will answer in a short time.
Is DataHub OSS Core free, and does that make it cheaper than Dawiso?
DataHub OSS Core is open source, but "free" shifts cost rather than removing it. Production self-hosting means a dedicated team standing up and maintaining Kafka, Elasticsearch, a metadata store, the GMS backend and a frontend, so the real total cost includes that infrastructure and operations work. DataHub Cloud pricing is opaque - the Enterprise tier is available on request with no published dollar amounts - so buyers can't budget either path upfront easily. Dawiso is managed in every deployment option with the full catalog on every plan, so there's no self-hosted ops burden and no "the operational features live in the paid tier" split.
How long does migration from DataHub take?
Dawiso reaches first use cases in 2-4 weeks, with a full named rollout (P3 Parks) in ~11 weeks. Migration from DataHub adds a metadata export through the OSS APIs and a mapping pass into Dawiso, which most teams complete inside one additional sprint. Dawiso open APIs make the export and re-ingest path straightforward.
Can we run DataHub and Dawiso in parallel during evaluation?
Yes. Point both at the same warehouse, give a small business team a 30-day shadow trial on Dawiso, and compare governance adoption side by side. Most teams notice the operations gap (managed catalog vs. self-hosted Kafka/Elasticsearch/GMS) within the first sprint.
When does DataHub still make sense?
If you have a platform team that wants to own and operate the stack, accepts the Cloud-positioned governance-lifecycle and SLA tooling, and has dedicated engineers to run Kafka, Elasticsearch, and GMS long-term, DataHub OSS Core can fit. If you want the catalog managed for you - one product, with the full catalog on every plan - that is the Dawiso case.
See why teams pick Dawiso over DataHub
Explore Dawiso in the real product, at your own pace. No sales call; a work email unlocks the full tour.
dawiso Winner 
DataHub Limited